| Message ID | 1093128c646b154a14d89321454f5361c0e616b4.1600854717.git.liu.denton@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | pull: do not warn when opt_ff is explicitly specified | expand |
Denton Liu <liu.denton@gmail.com> writes: > In d18c950a69 (pull: warn if the user didn't say whether to rebase or to > merge, 2020-03-09), `git pull` was taught to warn users if they > have `pull.rebase` unset or `pull.ff != "only"`. However, this warning > is a little too eager about happening. In d18c950a69, we added some tests to keep the warning working as intended, and the fact that this patch does not touch anything in the t/ directory means that its test coverage was not perfect. We'd need to add a test or two to make sure this new behaviour is protected against future breakages. Thanks. > If the warning is silenced by specifying `pull.ff = "only"`, as > instructed, the warning will arise again if the user runs > something like `git pull --no-ff`. However, the warning should not > happen as the user clearly knows what they're doing. > > Don't display the warning if opt_ff is explicitly set by a command-line > option given by a user. > > Signed-off-by: Denton Liu <liu.denton@gmail.com> > --- > builtin/pull.c | 5 ++++- > 1 file changed, 4 insertions(+), 1 deletion(-) > > diff --git a/builtin/pull.c b/builtin/pull.c > index 015f6ded0b..307b4b5d21 100644 > --- a/builtin/pull.c > +++ b/builtin/pull.c > @@ -83,6 +83,7 @@ static char *opt_commit; > static char *opt_edit; > static char *cleanup_arg; > static char *opt_ff; > +static int opt_ff_explicit; > static char *opt_verify_signatures; > static int opt_autostash = -1; > static int config_autostash; > @@ -344,7 +345,7 @@ static enum rebase_type config_get_rebase(void) > if (!git_config_get_value("pull.rebase", &value)) > return parse_config_rebase("pull.rebase", value, 1); > > - if (opt_verbosity >= 0 && > + if (!opt_ff_explicit && opt_verbosity >= 0 && > (!opt_ff || strcmp(opt_ff, "--ff-only"))) { > warning(_("Pulling without specifying how to reconcile divergent branches is\n" > "discouraged. You can squelch this message by running one of the following\n" > @@ -933,6 +934,8 @@ int cmd_pull(int argc, const char **argv, const char *prefix) > > if (!opt_ff) > opt_ff = xstrdup_or_null(config_get_ff()); > + else > + opt_ff_explicit = 1; > > if (opt_rebase < 0) > opt_rebase = config_get_rebase();
On Wed, Sep 23, 2020 at 3:56 AM Denton Liu <liu.denton@gmail.com> wrote: > > In d18c950a69 (pull: warn if the user didn't say whether to rebase or to > merge, 2020-03-09), `git pull` was taught to warn users if they > have `pull.rebase` unset or `pull.ff != "only"`. However, this warning > is a little too eager about happening. > > If the warning is silenced by specifying `pull.ff = "only"`, as > instructed, the warning will arise again if the user runs > something like `git pull --no-ff`. However, the warning should not > happen as the user clearly knows what they're doing. > > Don't display the warning if opt_ff is explicitly set by a command-line > option given by a user. Hi Denton, thanks for working on this! We can also assume that the user knows what they are doing and does not need a warning if they have run `git config --global pull.ff no`. So really, we can just get rid of the check for strcmp(opt_ff, "--ff-only") altogether and instead only check !opt_ff. Could you do that and add some more tests to t5521-pull-options.sh? -Alex
On Wed, Sep 23, 2020 at 12:25 PM Alex Henrie <alexhenrie24@gmail.com> wrote: > > Hi Denton, thanks for working on this! We can also assume that the > user knows what they are doing and does not need a warning if they > have run `git config --global pull.ff no`. So really, we can just get > rid of the check for strcmp(opt_ff, "--ff-only") altogether and > instead only check !opt_ff. Could you do that and add some more tests > to t5521-pull-options.sh? Actually, never mind, I'll send the patch myself. When I originally put in the warning I didn't stop to think that if the user is proficient enough to set pull.ff to any value, they really don't need a warning. -Alex
diff --git a/builtin/pull.c b/builtin/pull.c index 015f6ded0b..307b4b5d21 100644 --- a/builtin/pull.c +++ b/builtin/pull.c @@ -83,6 +83,7 @@ static char *opt_commit; static char *opt_edit; static char *cleanup_arg; static char *opt_ff; +static int opt_ff_explicit; static char *opt_verify_signatures; static int opt_autostash = -1; static int config_autostash; @@ -344,7 +345,7 @@ static enum rebase_type config_get_rebase(void) if (!git_config_get_value("pull.rebase", &value)) return parse_config_rebase("pull.rebase", value, 1); - if (opt_verbosity >= 0 && + if (!opt_ff_explicit && opt_verbosity >= 0 && (!opt_ff || strcmp(opt_ff, "--ff-only"))) { warning(_("Pulling without specifying how to reconcile divergent branches is\n" "discouraged. You can squelch this message by running one of the following\n" @@ -933,6 +934,8 @@ int cmd_pull(int argc, const char **argv, const char *prefix) if (!opt_ff) opt_ff = xstrdup_or_null(config_get_ff()); + else + opt_ff_explicit = 1; if (opt_rebase < 0) opt_rebase = config_get_rebase();
In d18c950a69 (pull: warn if the user didn't say whether to rebase or to merge, 2020-03-09), `git pull` was taught to warn users if they have `pull.rebase` unset or `pull.ff != "only"`. However, this warning is a little too eager about happening. If the warning is silenced by specifying `pull.ff = "only"`, as instructed, the warning will arise again if the user runs something like `git pull --no-ff`. However, the warning should not happen as the user clearly knows what they're doing. Don't display the warning if opt_ff is explicitly set by a command-line option given by a user. Signed-off-by: Denton Liu <liu.denton@gmail.com> --- builtin/pull.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)